Ten years ago, NASA began a targeted search for extraterrestrial intelligence at the Arecibo Observatory. Using technology, the Targeted Search System (TSS), largely developed by the SETI Institute, humanitys most sophisticated SETI program began. One year later, the search was terminated by a budget-conscious US Congress. Fortunately, the Institute secured the long-term loan of the equipment from NASA. Then with private funding, doubled the size and capability of the TSS, adding the unique ability to immediately test possible signals using two widely separated telescopes. Since 1995, under the name of Project Phoenix, Institute scientists and engineers have conducted the worlds most sensitive and comprehensive SETI program.

On November 27, Project Phoenix will inaugurate its New Search System (NSS) and begin a new phase in the search for ET. Based on a modular architecture and programmable integrated circuits, the initial NSS will have roughly the same search power as the old system but occupy less than one fifth of the space.

The compact size is a result of the advances in technology, especially in Field Programmable Gate Array (FPGA) chips. These versatile chips are widely used in the NSS to perform many signal processing functions such as digital filters and Fourier transforms. State-of-the-art chips can be expensive, but we have been fortunate to receive a large donation of chips from Xilinx. The FPGAs supply the processing power for three custom circuit boards that, along with a pair of commercial processor boards, reside in an ordinary PC. The PC with the added horsepower is called a Programmable Detection Module (PDM).

Each PDM handles all stages of the signal processing on a portion of the radio spectrum. For the initial NSS, each PDM will process 2 MHz. This is a big change from the TSS where large custom computer systems did only one stage of the processing on 10 MHz. (See figures 1 and 2.) The modular design of the NSS can be easily expanded to process more bandwidth as technology and funding permit.

The computing power of a PDM will allow it to complete processing an observation in about two-thirds the time of the TSS. This shortens the processing pipeline and with increased memory capacity, will allow follow-up analysis of the original data of a candidate ETI signal. That means the NSS must be identical at both sites since both sites must have the original data.

When expanded to process 100 MHz, the NSS will include more than 40 PDMs at each observing site. In order to schedule, configure, and monitor that many PDMs we need sophisticated control software running on fast and reliable computers. Sun Microsystems has donated four top-of-the-line Unix workstations with large flat panel monitors and disk archive systems to serve as the NSS control computers.

The NSS is expected to finish the Project Phoenix observing program in early 2005. By the middle of that year, the PDMs and their control computers will leave the Arecibo and Jodrell Bank observatories, and head for the Allen Telescope Array. There, along with a third 100 MHz system, they will be the first three "back end" processors for the ATA. Why three systems? The ATA provides multiple simultaneous "beams", or high angular resolution observations, within a two degree wide field of view. Searching three target stars simultaneously will not only triple the efficiency of the search, but should provide good mitigation of terrestrial interference. With the ATA, the NSS will carry out new search strategies with new capabilities through the end of the decade.